1007 — Combined Medication Adherence to Cardiovascular Risk Reduction Therapy

Bryson CL (Seattle HSR&D Center of Excellence (COE)), Liu CF (Seattle HSR&D COE), Sharp ND (Seattle HSR&D COE), Perkins M (Seattle HSR&D COE), Blough DK (Dept. of Pharmacy, University of Washington), Boyko EJ (Seattle Epidemiological Research and Information Center (ERIC)), Maciejewski ML (Durham VAMC)

Objectives:
To assess the prevalence of combined adherence to medications for cardiovascular (CV) risk reduction. Medical therapy is the cornerstone of risk reduction, but is ineffective if not taken. Maximal reduction in cardiovascular events results from adherence to all aspects of therapy that address CV risk factors, including antihypertensive (HTN), lipid (LDL), and diabetic (DM) therapy.

Methods:
A cohort of overlapping primary care patients (total N=8,749) taking medications for hypertension (N=8,404), hyperlipidemia (N=6,356), and diabetes (N=7,388) from 4 VA facilities was followed from 3/2002 to 12/2003. Adherence to medications was assessed based on a medication possession ratio (MPR) from electronic pharmacy records, defined as a proportion of days of drug available in a month. The MPR was constructed for each drug, and then a composite MPR was created for each of the 3 drug groups (hypertension, diabetes, and lipids). Patients were considered adherent to a drug group if their composite MPR > =80% for that month. Proportions adherent to 3/3 and 2/2 groups were created for the appropriate subgroups. Generalized estimating equations with a logit link were used to control for confounding and account for repeated measures in analyses.

Results:
The proportions of patients adherent to hypertension, lipid, and diabetic therapy on average each month were 50%, 72%, and 59%, respectively. Among 5,228 (60%) patients on therapy for all 3 conditions (DM/HTN/LDL), only 27% were adherent to all 3 therapies on average each month. Among 937 (11%) patients on only HTN/LDL therapy, 37% were adherent to both groups of medications on average each month. Age, number of drugs in regimen, month of followup, and disease severity were predictive of less overall adherence in adjusted analyses.

Implications:
Combined adherence to cardiovascular risk reduction medication was much worse than for each condition, indicating a large amount of potentially controllable risk. Patients who are adherent to medications for one condition may not be adherent to others.

Impacts:
An overall adherence gap gives a better estimate of unaddressed risk than focusing on adherence for single conditions. The magnitude of missed opportunity for CV risk reduction due to nonadherence may be large.